AI anime storytelling is crazy now
I used ChatGPT Image 2.0 to create an entire anime short film storyboard.
Then Seedance 2.0 turned it into a cinematic animated scene in minutes.
step by step tutorial with prompts: 👇
superintelligence warps the gravity of capitalism, trading any useful service with the machine world is the highest value thing any company can do. if SpaceX is a fab on the moon and a neocloud in orbit we’ll get the mars stuff as a minor side project
I cloned a premium landing page hero using https://t.co/kugrWQ38i8 + GPT-5.5… and honestly the result shocked me.
Inspired by https://t.co/LY7g2YyRcE 👀 What’s scary is not just the visual similarity.
AI now understands the *design DNA* of landing pages: spacing, typography, composition, glow, hierarchy, atmosphere…
I only added a https://t.co/ier5U6yoc5 background for this cinematic light beam effect ✨
This isn’t just “generate UI” anymore. It genuinely feels like AI understands creative direction.
And this is just the hero section 😅
$MSFT is currently trading below the 2D Bull Market Support Band, with $413 remaining my next key support level.
For now, I would like to see a reclaim of the 2D BMSB before turning more bullish.
Should you build the features you’re excited by?
Using excitement 𝘰𝘯𝘭𝘺, won’t make a great product. Must include “customer” & “vision/strategy.”
But ignoring excitement also isn't a path to greatness. Your taste matters,
Not a “balance,” but rather: “Both.”
https://t.co/62zncoaLI4
I’m launching a podcast and the first episode is out.
Main Branch is a show about how leading AI startups actually get built, deployed, and scaled. Each episode I’ll go deep with AI founders and early adopters covering their tech stack, research bets, inflection points, key hires, and topics top of mind for me and the founders I talk to day-to-day.
For the first episode I sat down with Arvind Jain @jainarvind, founder and CEO of @glean. Arvind took an idea most people had written off, enterprise search, and built it into one of the fastest-growing AI platforms out there, past $300M in ARR. He is now on a mission to expand human potential to do extraordinary work.
We get into:
• The early operating playbook, including sending 100 cold LinkedIn DMs a day
• The product arc from a search box to an agentic coworker with enterprise context
• Why he thinks the AI market is still 100x undersupplied, and why crowded markets shouldn't scare founders
• His case for how to compete when the labs move up the stack
• How AI has changed hiring (hint: you can't pass some of Glean's interviews without it)
More builders coming soon. Full episode below 👇
Two years ago I predicted fine-tuned SLMs would dominate financial workflows.
They haven't.
The models kept scaling, context engineering kept improving, and fine-tuning a model just to have a new frontier release erase that work became a recurring headache.
So I was wrong on the timeline. But I still think directionally this is correct, except that it's not about replacing frontier models for general reasoning, but rather focus on workflows that frontier models on their own can't serve.
I.e. tight latency loops, on-prem deployments, air-gapped data environments and unit economics.
Meanwhile, I think the labs have already/will won/win the long-running autonomous agent layer. Mostly because the harness of CC/Codex is becoming training data.
The opportunity is to build the tools these agents crave inside your domain, expose your product through MCP, and embed small models for the workflows that need them.
They can be complementary.
New post is up.
Introducing Vocs v2: a minimal docs framework designed for agents and humans.
Flexible docs that stay simple at the source, rich in the browser, and easy for agents to consume.
$ 𝚗𝚙𝚖 𝚒𝚗𝚒𝚝 𝚟𝚘𝚌𝚜
10 hires - $600K/yr saved.
Video editors. Graphic designers.
Retention managers. EAs.
Leandro Cabrini (@leacabrini), Founder at @weareproppel , on their work w/ @jonesroadbeauty:
"You can do faster iterations in-house without going crazy on cost."
Yesterday I got asked if it was possible to create a slide deck from an OpenBB dashboard.
So I connected @OpenAI Codex to @openbb_finance Workspace, via our recently launched Workspace MCP.
And pointed Codex to Alberto Gallini Open Portfolio application and asked:
> Create an investment-committee slide deck from the entire dashboard for us_stocks vs dj30, covering performance, sector attribution, holdings risk, factor/risk-model insights, and actionable portfolio recommendations. Use exact dashboard figures and make the deck polished, concise, and decision-oriented.
This 2min video is a compressed version of what happened in the next 20minutes. This was my first try at this, very happy with the output.
Next would be to have my design system / brand materials as part of the prompt so the slide deck has the formatting & style I'm interested in.
Anyway - this is very cool.